YOLO-World
YOLO-World copied to clipboard
about test prompt in inference
thanks for your nice work. i fine-tune the yolo-world model based on my custom dataset in coco-finetune mode , with the test.json = [ "A", "B", 'C" ]. when i run inference with image_demo.py after training, the detection results are bad. I check the code and delete the "[[' ']]" in image_demo.py, then the results are normal. now , i have some questions:
- whats the function of "[[' ']]" on yolo-world training and testing?